Automated Machine Learning on Azure
summary
Compares AutoML on Azure Machine Learning and Databricks with concrete results and tradeoffs: Databricks produces line‑by‑line notebooks, MLflow logging and feature store integration and returns a LightGBM best model runnable on Spark; Azure ML produces an interactive explainability UI, voting ensemble best model, and more problem types. Example run yields 42 models in 37 minutes and produced 0.924 vs 0.916 accuracy before test adjustment. Shows upload, validation split differences, pandas profiling data exploration, deployment options (ACI/Kubernetes/real‑time), and where manual feature engineering still matters.
AI-written summary · about
chapters (9)
- 0:00Introduction
- 1:35The case for AutoML
- 5:43Understanding AutoML
- 12:42Data preparation comparison
- 18:21Configuring AutoML runs
- 24:58Viewing experiment results
- 30:30Model analysis and comparison
- 39:12Model deployment
- 43:52Best practices and Q&A
related talks
58mDeep Dive into Azure Automated Machine Learning
48mMachine Learning in Azure Databricks
41mLeveraging Automated Machine Learning to enable anyone to develop Machine Learning solutions
40mCloud-based Machine Learning Offerings - Beginner's Perspective
43mA lap around Azure Machine Learning services
48mStreamlining your Machine Learning projects with Azure Machine Learning
01h 05mAzure Machine Learning for the Developer
51mTaking Models to the Next Level with Azure Machine Learning